From: Alexander Monakov <amonakov@ispras.ru>
To: gcc-patches@gcc.gnu.org
Cc: Alexander Monakov <amonakov@ispras.ru>,
Richard Biener <richard.guenther@gmail.com>
Subject: [PATCH 2/3] tree-cfg: do not duplicate returns_twice calls
Date: Fri, 14 Jan 2022 21:20:46 +0300 [thread overview]
Message-ID: <20220114182047.6270-3-amonakov@ispras.ru> (raw)
In-Reply-To: <CAFiYyc3h+fNQ4KdyRWBeVM=NESqSWbbJNB8iGctx_UZ5BVVyfA@mail.gmail.com>
A returns_twice call may have associated abnormal edges that correspond
to the "second return" from the call. If the call is duplicated, the
copies of those edges also need to be abnormal, but e.g. tracer does not
enforce that. Just prohibit the (unlikely to be useful) duplication.
gcc/ChangeLog:
* tree-cfg.c (gimple_can_duplicate_bb_p): Reject blocks with
calls that may return twice.
---
gcc/tree-cfg.c | 7 +++++--
1 file changed, 5 insertions(+), 2 deletions(-)
diff --git a/gcc/tree-cfg.c b/gcc/tree-cfg.c
index b7fe313b7..a99f1acb4 100644
--- a/gcc/tree-cfg.c
+++ b/gcc/tree-cfg.c
@@ -6304,12 +6304,15 @@ gimple_can_duplicate_bb_p (const_basic_block bb)
{
gimple *g = gsi_stmt (gsi);
- /* An IFN_GOMP_SIMT_ENTER_ALLOC/IFN_GOMP_SIMT_EXIT call must be
+ /* Prohibit duplication of returns_twice calls, otherwise associated
+ abnormal edges also need to be duplicated properly.
+ An IFN_GOMP_SIMT_ENTER_ALLOC/IFN_GOMP_SIMT_EXIT call must be
duplicated as part of its group, or not at all.
The IFN_GOMP_SIMT_VOTE_ANY and IFN_GOMP_SIMT_XCHG_* are part of such a
group, so the same holds there. */
if (is_gimple_call (g)
- && (gimple_call_internal_p (g, IFN_GOMP_SIMT_ENTER_ALLOC)
+ && (gimple_call_flags (g) & ECF_RETURNS_TWICE
+ || gimple_call_internal_p (g, IFN_GOMP_SIMT_ENTER_ALLOC)
|| gimple_call_internal_p (g, IFN_GOMP_SIMT_EXIT)
|| gimple_call_internal_p (g, IFN_GOMP_SIMT_VOTE_ANY)
|| gimple_call_internal_p (g, IFN_GOMP_SIMT_XCHG_BFLY)
--
2.33.1
next prev parent reply other threads:[~2022-01-14 18:21 UTC|newest]
Thread overview: 25+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-12-13 14:25 [RFC PATCH] tree-ssa-sink: do not sink to in front of setjmp Alexander Monakov
2021-12-13 14:45 ` Richard Biener
2021-12-13 15:20 ` Alexander Monakov
2021-12-14 11:10 ` Алексей Нурмухаметов
2022-01-03 13:41 ` Richard Biener
2022-01-03 16:35 ` Alexander Monakov
2022-01-04 7:25 ` Richard Biener
2022-01-14 18:20 ` Alexander Monakov
2022-01-14 18:20 ` [PATCH 1/3] " Alexander Monakov
2022-01-17 7:47 ` Richard Biener
2023-11-08 9:04 ` Florian Weimer
2023-11-08 10:01 ` Richard Biener
2023-11-08 13:06 ` Alexander Monakov
2022-01-14 18:20 ` Alexander Monakov [this message]
2022-01-17 8:08 ` [PATCH 2/3] tree-cfg: do not duplicate returns_twice calls Richard Biener
2022-07-12 20:10 ` Alexander Monakov
2022-07-13 7:13 ` Richard Biener
2022-07-13 14:57 ` Alexander Monakov
2022-07-14 6:38 ` Richard Biener
2022-07-14 20:12 ` Alexander Monakov
2022-07-19 8:40 ` Richard Biener
2022-07-19 20:00 ` Alexander Monakov
2022-07-13 16:01 ` Jeff Law
2022-01-14 18:20 ` [PATCH 3/3] tree-cfg: check placement of " Alexander Monakov
2022-01-17 8:12 ` Richard Biener
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20220114182047.6270-3-amonakov@ispras.ru \
--to=amonakov@ispras.ru \
--cc=gcc-patches@gcc.gnu.org \
--cc=richard.guenther@gmail.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).